RMPE: Regional Multi-person Pose Estimation

Benchmark Model Rank Results
2d-human-pose-estimation-on-ochumanRMPE#7Test AP: 30.7Validation AP: 38.8
keypoint-detection-on-cocoAlphaPose#11Test AP: 73.3FPS: 23
keypoint-detection-on-coco-test-devAlphaPose#15APL: 81.5
keypoint-detection-on-mpii-multi-personAlphaPose#1mAP@0.5: 82.1%
keypoint-detection-on-ochumanRMPE#6Test AP: 30.7Validation AP: 38.8
multi-person-pose-estimation-on-coco-test-devRMPE#12AP: 61.8APL: 67.6APM: 58.6AP50: 83.7AP75: 69.8
multi-person-pose-estimation-on-crowdposeAlphaPose#19mAP @0.5:0.95: 61.0AP Easy: 71.2AP Medium: 61.4AP Hard: 51.1
multi-person-pose-estimation-on-mpii-multiAlphaPose#1AP: 82.1%
pose-estimation-on-coco-test-devRMPE++#25AP: 72.3AP50: 89.2AP75: 79.1APL: 78.6APM: 68.0
pose-estimation-on-coco-test-devRMPE#40AP: 61.8AP50: 83.7AP75: 69.8APL: 67.6APM: 58.6
pose-estimation-on-ochumanRMPE#15Test AP: 30.7Validation AP: 38.8